Armin Laschet gets the most hateful comments on the net

An investigation by Hateaid, which specifically dealt with verbal attacks against the candidates for chancellor and other leading personnel of the parties, approaches the extent of hatred and agitation on the short message service Twitter alone.

Hateaid took a closer look at the volume of publicly posted hateful comments, especially against the three top candidates for the Federal Chancellery for the period from August 7 to September 7, 2021. Using the Perspective API Tweets were evaluated and categorized as hate speech if the AI ​​algorithms used were able to determine at least 85 percent potential for insults according to the API definition.

In this way, they found around 40,000 tweets on the three top candidates from the CDU, SPD and Greens alone that could be classified as hateful comments. The noticeable thing about it: The agitation is not evenly distributed between the applicant and her two competitors.

Rather, the CDU candidate Armin Laschet is much more affected by expressions of hatred and insults than Olaf Scholz from the SPD or Annalena Baerbock from the Greens. Of the almost 40,000 hate tweets, 27,476 are directed against Laschet alone. There are 6,690 hate tweets against Scholz and 5,524 against Baerbock.

It is also true, however, that the basic amount of tweets that deal with the individual person is just as significantly different. Hateaid has collected over a million tweets over the month.

Laschet was mentioned around 660,000 times in tweets during the period under review. Scholz got around 280,000 mentions and Baerbock around 230,000 mentions. Nevertheless, in relation to the mentions, Laschet remains clearly the most affected with around four percent hate tweets. Scholz and Baerbock each come to a little over two percent.

FDP leader Christian Lindner also had to accept that he was insulted 1,599 times in around 57,000 tweets. The representatives of the parties at the edge of the spectrum were comparatively less affected. Janine Wissler from the Left was insulted in 283 of around 9,300 tweets, Alice Weidel from the AfD in 684 of around 27,000 tweets.

In the next step, the AI ​​used by Hateaid assessed the criminality of the postings and, based on a small sample, came to the conclusion that around seven percent of the tweets classified as offensive and / or offensive also constitute a criminal offense. In the election campaign, the limits of what top politicians in particular have to endure in the name of freedom of expression before it turns into a criminal offense are shifting.

In addition, Hateaid has tried to quantify how high the proportion of hate tweets from the right spectrum is. Here the organization has curated a portfolio with around 10,000 accounts and evaluated postings from it. It turned out that the amount of hate tweets from the right increases the more the politicians’ programmatic positions are to the left.

In the case of Janine Wissler, 36 percent of all hateful comments according to Hateaid’s reading are rightly motivated, with Annalena Baerbock it is 25 percent, with Olaf Scholz 21 and with Armin Laschet a further seven percent. Conversely, this also means that even with Janine Wissler 64 percent of all hate tweets could not be assigned to the right spectrum, with Armin Laschet even 93 percent. It would be interesting to see how the entire political and other social spectrum is distributed among the spread of hate. Hateaid just bothered to look in the right corner. The problem, however, must be much bigger.

In terms of content, the hate comments from all candidates according to Hateaid’s findings are extremely similar. While Armin Laschet was often insulted with the terms “failure”, “fress” or “idiot”, Olaf Scholz follows at a certain distance, who also had to be called a “failure”, but also as a “puppet” or “devil” . Annalena Baerbock also frequently used the term “failure”, but the Green candidate was also referred to as “Stasi” or “informer”. The term “failure” was also found in all other candidates with varying frequency, with women additionally using the term “bitch” and in the case of Alice Weidel “Nazi bitch”.

Hateaid presents the entire report, which also contains evaluations of more than 500 Telegram channels and groups, under the title “Hate as an occupational risk – digital violence in the election campaign” and this link as a PDF ready for download.